[offsides]

I've mulled over Ogre campaigns for a long time, and I think they keys to making it work are as follows:

1) While there may be minor benefits/penalties for winning/losing a prior battle, each individual battle needs to be self-contained. For example, the starting forces for each battle should be (mostly if not completely) independent from what survived the prior battle. There may be one or 2 things that depend on previous outcomes (e.g., if you saved a particular objective you might have 1 or 2 extra units but nor a whole division, or maybe you get reinforcements a turn sooner), but nothing that's highly unbalancing.

2) Battles should be about objectives and not just annihilation. While blowing up your opponent to the last is all fun and good, it makes for a bad campaign setup. Both because losing a few of those kinds of battles can be depressing for the loser, and because it doesn't really advance a campaign in meaningful ways. In any long campaign, troops are going to try to minimize losses and either secure their objective or make a strategic withdrawal. So make campaign battles the same - both sides want to accomplish their objective(s) if possible, but if not they want to live to fight another day. Depending on the map and scenario, both sides may have objectives that allow them to "advance" should one side achieve them and the other doesn't, or it may be a case where one side can only "hold" their ground and prevent an advance while the other side can advance should they reach their goals. If you play the latter type of scenarios, make sure you alternate which side is trying to advance and which side is trying to hold, so that one side can't just steamroller over the other.

3) Big battles are fun, but small battles win the war. It's great to come up with big battle scenarios, but they take too long and too many wins on one side can smash the campaign quickly. Stick to mostly smaller skirmishes, saving the big battles for special occasions. Perhaps each of the smaller battles can influence the starting setup/reinforcements of a later big battle, with the caveats listed in #1 about not unbalancing things too badly.

4) Finally, don't make it too big. Yes, a global war is gonna take a huge amount of time to simulate, but you don't have to do it. Keep the total campaign to a small enough size that you can reasonably expect to finish it within 10-12 battles. That's not to say it won't finish faster if one side gets all the luck, nor do you have to stop cold when it's over (you can always roll into a new campaign right away). But it's not worth spending months planning a huge campaign that you barely scratch the surface of. That's not to say that I wouldn't do that at all (because I totally would :P), but isn't it better to play the campaign rather than just planning it? :)

I'm still percolating on my 'Tug-O(gre)-War' idea that essentially uses the campaign concept on a single Ogre battlefield setup as well, which has influenced some of these ideas. But maybe this discussion will kick that loose enough to finally write it up...